Log in

View Full Version : pocketnow.com Review Linksys' Wireless-G CF Card


Darius Wey
01-26-2005, 04:00 PM
<div class='os_post_top_link'><a href='http://www.pocketnow.com/index.php?a=portal_detail&t=reviews&id=557' target='_blank'>http://www.pocketnow.com/index.php?...=reviews&id=557</a><br /><br /></div><i>"Wireless connections can be both a blessing and a curse. Once you've accustomed yourself to having WiFi, going to a location without it is quite a letdown. And if you have WiFi available, one has to determine how much of an effort they are willing to go through in order to connect. Even after all that effort, you may only connect at half the speed you anticipated! Linksys, with the release of their Wireless-G CompactFlash Card (the first of its kind), hope to at least make some of these issues a moot point (at least for your PDA). But is a Wireless-G card really worth having on a PDA? Can your wireless woes be healed with a single CF card?"</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/wey-20050126-WirelessG.jpg" /><br /><br />Tim from pocketnow.com had a chance to play around with the Linksys Wireless-G CF Card. Think 802.11g makes a difference? I guess you'll have to read the review to find out. But I'm hoping to see SD versions of these 802.11g cards appear sometime soon. ;-)

Deemo
01-26-2005, 05:04 PM
I have always been under the assumption that the PPC bus speed would limit any speed gains that G had to offer.
Is this incorrect?

one504
01-26-2005, 05:25 PM
I have always been under the assumption that the PPC bus speed would limit any speed gains that G had to offer.
Is this incorrect?

No, I think you're right - that's why the file transfer times were the same (the file has to be written to storage), but web pages (in RAM) were much faster. Just a guess, though.

ctmagnus
01-26-2005, 05:47 PM
afaict, the sole advantage to this card, when used in a Pocket PC, is that it won’t drag down other G users. Of course it can be used in laptops/desktops as well.

Menneisyys
01-26-2005, 05:51 PM
"In the file transfer test, I connected to a personal server and downloaded a 300k .mp3. ... Strangely, the 300k .mp3 file transferred within 24 seconds on both devices. The Linksys card was faster by 1 second during one transfer, but then was tied the second time through."

24 seconds for a 300 kbytes file is VERY slow - there must have been a serious problem in the setup. vxFTP, for example, downloads stuff to main memory via 802.11b with around 300 kbytes/s, even in an iPAQ 2210. That is, using vxFTP, the file would have arrived in a second (or less), if you subtract the latency due to opening data connection ports (some tenths of seconds at most, even on a PPC).

The review should have mentioned the FTP client. I assume it was the Resco Explorer FTP client, downloading to a storage card, because no other FTP clients / setups are so slow. This also explains why the 802.11b / 802.11g clients downloaded the file with the same speed, while the Tom's Hardwre people, in their review at http://www.tomsnetworking.com/Reviews-177-ProdID-WCF54G.php (see also http://www.pocketpcthoughts.com/forums/viewtopic.php?t=35287 ), did see some speed difference.

See http://www.pocketpcthoughts.com/forums/viewtopic.php?t=35119 for more on FTP client speeds.

Sven Johannsen
01-26-2005, 05:52 PM
I have always been under the assumption that the PPC bus speed would limit any speed gains that G had to offer.
Is this incorrect?
I think that factors in as well.

There were just too many variables on the tests to make me think there was anything that could be specifically attributed to the a/g difference.

Downloading a file from the local network is the only thing you could reasonably control the variables. But then you'd want a B card and a G card in the same PPC to test the change. Preferably two linksys cards too.

Testing WEB downloads, unless you have better than a 54Mb connection to the internet would seem sort of fruitless. If you have a 1-1.5Mb connection to the net (cable, DSL, T-1, etc.) that is the local bottleneck. If the info doesn't come into the AP at more than 1.5Mb, then whether it gets from there to the PPC at 11Mb or 54Mb is somewhat irrelevent.

I have to assume that care was taken to ensure the AP was operating in G only modes to test that, and B only modes to test that. I understand that mixed clients on an A/G AP will affect the throughput of the G clients adversely, but I haven't seen a definitive quantitative review.

Menneisyys, where does the term ftp appear in the review? It just says he connected to the server and downloaded the file. That could easily be done just by openning a share and copying while watching the second hand on your watch ;)

Menneisyys
01-26-2005, 05:55 PM
that's why the file transfer times were the same (the file has to be written to storage)

Nope - it's just that prolly Resco Explorer's FTP client has been used, whch is dog-slow when writing to storage cards and this is why the results were exactly the same. The HTTP test, on the other hand, does show that the g card (as with the Tom's Hardware benchmarks) is indeed faster even in current PDA's.

Menneisyys
01-26-2005, 06:02 PM
BTW, another CF card not having an external aerial socket... :(

Menneisyys
01-26-2005, 06:08 PM
Menneisyys, where does the term ftp appear in the review? It just says he connected to the server and downloaded the file. That could easily be done just by openning a share and copying while watching the second hand on your watch ;)

Yes, it's not said what kind of protocol and client have been used. Not that it would be important - the resulting 12.5 kbytes/s speed is clearly the result of the protocol / the client, and not the PDA architecture / the Wi-Fi cards. Over 802.11b, using plain FTP and a decent FTP client (vxFtp), 30 times more speed is achievable. I think using the 802.11g card under these circumstances would have resulted in even more speed - around 400-500 kbytes/s.

WyattEarp
01-26-2005, 06:54 PM
It's good to see that there are benefits to having 802.11g in your PDA disspite the limitations (bus speed, actual connection speed, network speed, etc.) But this is what you deal with every day you go online; sometimes pages load fast sometimes they load slow. But as long as the capability of fast loading is there it's all good :D .
But I won't truely be happy until I get it built-in to my PDA though... Is anyone listening... It's about time... :|

hawkeye
01-26-2005, 10:42 PM
The point of this card is not faster access for the PDA!!

The point is not to drag a G only network down to B speeds.

For instance, I have a G only network at home. The access point is set to only accept G connections. I have a 2215. If I wanted wifi on my PDA my only option was a B card, which would have required me to make my G access point accept B cards which would have forced my G network to operate at lowly B speeds.

With this new Linksys card I can add wifi to my 2215 and not degrade the performance of my other wifi PCs.

That's the only purpose of this card as I see it.

Any thoughts??

phs
01-27-2005, 02:28 AM
The point is not to drag a G only network down to B speeds.

Wireless-B cards cost only $32. This is one instance when concern for others can cost ya.

Menneisyys
01-27-2005, 09:26 AM
The point of this card is not faster access for the PDA!!


It does provide (slightly? Considerably? Dunno, because the file trasfer test used the wrong protocol/client, so no real speed difference could have been measured) some speed advantage too, according to the Tom's Hardware test.

fishfuzz
01-27-2005, 04:07 PM
that's why the file transfer times were the same (the file has to be written to storage)

Nope - it's just that prolly Resco Explorer's FTP client has been used, whch is dog-slow when writing to storage cards and this is why the results were exactly the same. The HTTP test, on the other hand, does show that the g card (as with the Tom's Hardware benchmarks) is indeed faster even in current PDA's.

Resco wasn't used for the test. It was a connection via IE directly to the .mp3. Both .mp3 tests were stored in the My Documents directory.

fishfuzz
01-27-2005, 04:16 PM
"In the file transfer test, I connected to a personal server and downloaded a 300k .mp3. ... Strangely, the 300k .mp3 file transferred within 24 seconds on both devices. The Linksys card was faster by 1 second during one transfer, but then was tied the second time through."

24 seconds for a 300 kbytes file is VERY slow - there must have been a serious problem in the setup.

Actually, the test was not designed to demonstrate maximum throughput (which could have had the results marred by any number of factors). The test was configured to test real life day-to-day situations (ie ones were you wouldn't normally have an excellent ftp client). The .mp3 was stored on an average server with average bandwidth to balance the overall real world results.

I, for one, like max throughput numbers, but I can get these from the manufacturer. What they don't talk about (or include in their marketing materials) is how well the unit works when it's installed on _your_ system.

Menneisyys
01-27-2005, 04:20 PM
I, for one, like max throughput numbers, but I can get these from the manufacturer.

Manufacturers tend to li... uhm, tell us overly optimistic "results" :) This is why I'd really be interested in what the real preformance of a Wi-Fi unit is.

fishfuzz
01-27-2005, 04:24 PM
True, but either way the results will be disappointing when you get the unit into your environment unless the original tests were designed to match similar settings.

Menneisyys
01-27-2005, 04:33 PM
True, but either way the results will be disappointing when you get the unit into your environment

Well, I'd say it depends. For example, even at very weak signals, built-in/add-on Wi-Fi units on the PPC platform (because of the PDA hardware bottleneck) don't deliver much worse speed than in an ideal case (except for large, seconds-long, occassional drop-outs or even disconnects). I have never encountered much less speeds than 250 kbytes/s from intranet downloads/file access via vxFtp, however poor the signal was.

Of course, it's intranet only; accessing the internet all depends on the speed of the internet connection / the speed of the connection to the host. I really missed this additional information from the review because, as with the Tom's Hardware Linksys 802.11g test, I thought you wanted to measure the maximal throughput of the card and, therefore, also used a file server on the local intranet.